November 03, 2010
The three R's - repetition, repetition, repetition -- helps establish control of your dog.
When training young Blackie, you've surely sometimes wished you could explain a command to him in simple English. For example: "Look, Deadgrass, when I say 'Heel,' you just saunter along beside me. Trust me; this'll come in handy later on."
Or, "Goldilocks, when I say 'Hold,' please keep this dummy in your mouth until I say 'Give.' Is that asking too much?" And so on.
But dogs can't understand ordinary conversation (which is sometimes a blessing). Consequently, we must instill control training through "conditioning," compliments of Pavlov and his salivating pooch. Conditioning has three elements: stimulus, response and results. The stimulus is a command word (Sit, Heel, Come and so forth); the response is the dog's reaction, correct or incorrect and the results vary with the dog's reaction. When he responds correctly, we reinforce that response by rewarding him. When he responds incorrectly, or not at all, we dissuade him from such behavior by punishing him. After sufficient repetitions, the dog learns to react properly and to avoid reacting improperly to the command. Clearly, the cornerstone of conditioning is repetition. Thus, the "three R's" of control training are: Repeat, Repeat, Repeat. Of course, we should not repeat anything too often in a given training session, lest we damage the dog's spirit.
When the dog responds properly and is thereby rewarded, he's in "Trial and Success" mode. When he responds improperly and is thereby punished, he's in "Trial and Error" mode. The former encourages proper responses by motivating positively. The latter discourages improper responses by motivating negatively. The positive motivation of "Trial and Success" promotes style. The negative motivation of "Trial and Error" promotes reliability. To develop a retriever that is both stylish and reliable, you must use both, but in the proper sequence and mixture. Too much of the former will adversely affect reliability. Too much of the latter will adversely affect style.
You should start each command out with "Trial and Success." You give the command, and then help your dog obey properly, and finally you reward him. With enough repetitions, he'll learn what you expect of him and do it happily. After he understands the command and obeys it without assistance, you should add enough "Trial and Error" training to convince him that he must obey every time. Here you give the command in circumstances (distractions and so forth) that tempt him to disobey. When he does, you punish him appropriately. With enough repetitions in enough distracting situations, he'll learn that he must obey, even when he thinks he has a better idea. Thereafter, you should mix the two in whatever proportion your individual retriever requires to maintain both style and reliability.
Phase 1: Trial And Success
In this phase, you use positive techniques to teach your dog what you expect of him on a given command. To do this, you rely on training exercises that preclude failure and ensure success, with its subsequent reward. Let me give you a couple of examples.
When you start steadying your retriever, to prevent him from making a mistake (breaking), you physically restrain him until you say his call-name (or whatever you say to send him to retrieve). The retrieve itself is his primary reward, with the icing being a little praise when he delivers to you. Since successful retrieves matter so much, you should do all phase-one steadying in coverless fields, where he'll always find the bird or dummy easily. Enough repetitions of successful "restrained" retrieves will condition him to wait for his call-name before leaving to retrieve. In his early puppy retrieves you can prepare your dog for this steadying process by saying his name as he takes off after each just-thrown puppy dummy. Toss the dummy, and as he leaves to chase after it, say his call-name. If you do this consistently, he'll associate his call-name with leaving to retrieve. Then, when you start the formal steadying process, this mental association will kick in and help get the new concept across.
In teaching a retriever to "line" to blind retrieves, you should first mark the dummy pile with something visible (white cone, white flag, white bucket or white whatever). That way, when you give him your blind retrieve sequence (Dead bird€¦Good€¦Back or whatever) he has something to run to in order to find a dummy. After enough repetitions, when you give him your blind retrieve sequence, he'll take off and run straight ahead, even without the visual aid. Here too the retrieve is the primary reward, but you should also praise him.
And so on. For every control command, you can find a way to instill the proper response through totally positive drilling. In that way, your dog not only learns what you expect, but he learns to enjoy doing things your way.
Totally Positive Reinforcement
Some obedience trainers claim that this "Trial and Success" phase alone will completely train a dog, that punishment has no place in "civilized" dog training. They call their approach, "totally positive reinforcement." This sounds good, but it just doesn't work. Let me give you an example. Let's say you've taught your retriever the Sit-whistle with totally positive reinforcement. When you toot the Sit-whistle, he sits, and you immediately heap unimaginable pleasantries upon him. He so looks forward to these rewards that his rump hits the turf in mid-whistle, every time. How nice! Now take him pheasant hunting and blow the Sit-whistle while he's chasing a just-flushed hen. If he's any kind of a retriever at all, he'll opt to forgo whatever delights you may have in store for him so that he can continue chasing that hen. However, had you followed your "Trial and Success" training with appropriate "Trial and Error" reinforcement, his fear of punishment would stop him when his anticipation of rewards alone wouldn't even slow him down. After you've convinced him that "whistle refusals" (as these are called in polite retriever circles) always bring on the "bumble-bees" (jolts from the e-collar), he'll stop reliably under the most trying circumstances.
Phase 2: Trial and Error
For reliability in each command, after your dog understands what he's supposed to do, you should put him into situations that tempt him to disobey. That way, he'll give you opportunities to instill into his sometimes-wayward heart a "holy fear" that'll strongly discourage disobedience. After ample positive reinforcements, a little punishment won't dampen his spirit, especially if mixed in with plenty of positives.
Here are some examples. After your dog heels properly, you should introduce various distractions while he's heeling. Have family members walk around in the yard. Have someone bring in another dog, or perhaps better yet, a cat. Heel him around a shopping mall. Whenever he strays from your side, give him a sharp leash correction to bring him back to the heel position, and then pra
ise him for again being where he belongs. Introduce similar distractions to the Come-whistle, with him on a check cord so you can control him. Later, when he's returning with a dummy or bird, toss another dummy or bird off to either side of him, and insist that he ignore it, at least until he has delivered his current "payload." Here, too, for control, you should start out with him on a check cord before graduating him to free association.
You can use this two-phased approach in every form of control training. Phase one instills proper canine behavior and a positive attitude. Phase two convinces the dog that a command is not a polite request that he can safely ignore. To make phase two effective without damaging your dog's spirit, you must control the frequency of error in each session by introducing distractions gradually, not in quantum leaps. You want your dog to fail sometimes, so you can correct him and thereby teach him to avoid misbehavior. But, you don't want too many failures in a row, nor too high a percentage of failures overall. Either will damage the positive attitude you've developed in phase one.